Unlike AI tools that stop at interview + transcript, Mira is a full AI researcher — plans studies, recruits globally (100M+ panel, 120 countries), runs dynamic interviews with intelligent probing, and uniquely captures what participants say AND feel via real-time facial coding, voice emotion AI, and webcam eye tracking. Extracts themes, generates insights, and produces research reports automatically. 17 patents. 70+ languages. Trusted by Unilever, Nestlé and 150+ global brands. $25M Series B.
Forum Threads
The say/feel gap in research; how big is the problem actually, and what do you do about it?
There's a well-documented phenomenon in consumer and user research that most practitioners know intuitively but rarely name directly: people don't report their experience accurately.
Not because they're dishonest. Because self-reporting is hard. In the moment of an interview, participants are performing a version of themselves. They round off hesitation. They describe their behaviour more charitably than it actually was. They say "yes, I'd probably use this" when what they felt was closer to "maybe, under the right circumstances, if the price were different." The social pressure of being in a conversation, even with an AI, shapes what gets said.
This is what we call the Say-Do Gap. The distance between what someone tells you and what they actually feel or do.
It shows up everywhere. In concept tests where participants say they love a product they'd never buy. In usability sessions where someone says "this makes sense" while visibly struggling. In brand perception studies where stated attitudes don't match purchasing behaviour.
What do transcripts miss that sends you back to the recording?
Nine years of working closely with insights teams across CPG, BFSI, and tech has given us a consistent observation: the transcript is where researchers start, but the recording is where they go when the transcript isn't enough.
They go back to hear how someone said something. To check whether the pause before an answer was genuine uncertainty or just thinking time. To see whether the person's expression changed when they described a feature they claimed to like. To catch the moment when engagement visibly dropped when they stopped leaning forward, when their eyes moved away from the screen even while their words stayed polite and positive.
That gap between the transcript and the recording is where a significant amount of research insight lives. And it's also where the most time gets consumed in manual analysis. Researchers who run 20 or 30 interviews know what it's like to sit through hours of footage looking for the three or four moments that actually matter.
This observation that the transcript captures what was said but not how or with what underlying feeling is fundamentally what drove us to build the emotion and behavioural layer into Mira. Not to replace researcher interpretation, but to give researchers a way to find the right moments faster.
I'm genuinely curious: what's the signal you find yourself wishing you had after you've finished going through a set of interviews? What consistently drives you back to the recording that the transcript alone doesn't give you? And have you found any workarounds annotation practices, tagging systems, collaborative review setups that help you capture more of what the transcript loses?
What would actually make you trust an AI moderator's findings?
We launched Mira on Product Hunt today, and the conversations in the comments have been the most interesting part of the whole day.
Researchers asked about probe neutrality whether the follow-up questions an AI asks mid-interview can lead a participant, rather than uncover them. They asked about cultural calibration, whether emotion models trained largely on Western data can accurately read a participant in Jakarta or Nairobi. They asked whether the say/feel mismatch gets surfaced as raw evidence or quietly resolved into a single confidence score.
These are serious questions. And they made me realize something: the bar for trust in AI-moderated research is fundamentally different from other AI tools.
If a writing assistant gets something wrong, you catch it before you publish. If an AI coding tool hallucinates, your tests fail. But if a research tool misreads how participants felt during a concept test, and that feeds into a product decision, the error is invisible. By the time the product ships and the market responds, the research moment is long gone.
I think the difficult part is whether those signals truly reflect what someone is feeling at that moment. Sometimes the face expression, eye movement can mean different things depending on the person and the context. How do you validate that the emotional signals are accurate ?
Mira
@reda_roqai_chaoui That's the right challenge to raise, and honestly, it's why we don't treat any single signal as ground truth.
A raised eyebrow or gaze shift can mean genuinely different things depending on the person, culture, and context we agree completely. That's precisely why MIRA fuses multiple channels (facial, vocal, speech patterns) rather than scoring emotion off one modality. When signals agree across channels, confidence in the read is high. When they conflict, that's flagged as ambiguous rather than forced into a clean label.
running facial coding and eye tracking across 120 countries means dealing with wildly different consent and biometric data laws (BIPA in Illinois, GDPR in the EU, etc). is that handled per-region automatically or does it fall on the researcher to configure what's legal where they're recruiting from?
Mira
@omri_ben_shoham1 It is handled automatically by our platform—researchers never have to manually configure local legal frameworks. We are fully GDPR compliant (and SOC 2 Type II certified), seamlessly managing global biometric laws like BIPA across 120+ countries. First, we require explicit consent from every single respondent before initiating eye tracking and facial coding, ensuring universal legal alignment. Second, we leverage Edge Processing to calculate gaze and facial metrics locally in real time, meaning sensitive raw video streams are never transmitted across borders or stored centrally. You simply launch your study, and our infrastructure ensures every session is legally watertight and privacy-first!
the edge processing point is the one that actually reassures me, not shipping raw video across borders closes off a whole category of risk. the part I'd still want to see before trusting it fully is what the consent flow actually looks like from the respondent's side, is it a real explanation of what's being captured or a buried checkbox they click through to get to the questions
Mira
@omri_ben_shoham1 When it comes to building trust. To ensure the consent flow is never a 'buried checkbox,' we use a strict, two-step opt-in process before a respondent ever enters the study:1. We integrate with third-party panels where respondents must explicitly opt in to webcam and microphone access at the profile level. Only participants who have already consented to webcam-based studies receive the invite link. 2. In-App Instruction & Consent Screen (Second Gate) Even with panel-level pre-screening, we do not assume consent. When a respondent clicks the study link, they arrive at a dedicated instruction screen before the test begins.
two gates is a solid answer, that actually addresses it. one more thing I'd wonder about as a respondent: can I revoke consent mid-session if I get uncomfortable partway through, or is it all-or-nothing at the start?
How do i verify the accuracy of Facial coding and eye tracking?
Mira
@jitender_pankaj1 Our algorithms are trained and validated with millions of data points we collected over the years with our in house data collection and tagging platform. We ensure each tag has high inter rater reliability before accepting it in our dataset. Our eye tracking algorithms are frequently tested against the data from a physical eye trackers.
How does the pricing actually scale for smaller teams or solo researchers who don't need the full enterprise setup, and is there any way to try it before committing?
Mira
@barwcts The best starting point is the free study, no commitment, no credit card. Mention PH when you book: Book a Demo: See Entropik Decode in Action
For pricing beyond that, it scales based on study volume which can be purchased in smaller blocks. Happy to walk through what makes sense for your setup. Drop a note or book a quick call.
"Said yes, looked confused" catching that in real time instead of
buried in hour 3 of a recording is the kind of detail that makes me
trust the rest of the data way more.
Mira
@ulykbek11 When our models see the face or voice disagreeing with the transcript, the AI doesn't guess it just uses the live agent to ask the user: "I noticed a quick pause there what were you thinking?" That real-time feedback loop is what turns messy, subjective signals into data you can actually build on.
Respect for not hand-waving that, most launch threads would have. One thing I'd add: per-frame confidence is the model scoring its own certainty, so it won't catch systematic bias. A model can be high-confidence and wrong the same way across a whole population and never flag it. The only check I trust is human-coded ground truth sampled per region, which is painful to collect. Which regions have you actually validated against local human coders versus carried over from the base model?
Mira
Head of Sales at Decode here.
The conversation I have most with research and insights teams: "Our studies take too long, and leadership does not trust the findings."
Both problems have the same root cause. The tools being used only capture what people say, and analysis is manual. A 20-participant qual study typically takes 6-8 weeks from setup to report, arriving too late to influence the decision it was commissioned to support.
Mira compresses that timeline. Study setup with templates takes minutes. Recruitment from a 100M+ global panel is built in. Transcripts, themes, and reports are generated automatically. Emotional signal adds defensibility to findings.
If you run an insights function and want to understand what this looks like for your team's specific workflow, feel free to ask below.